KR830002643B1 - Audio frequency noise suppression circuit - Google Patents

Abstract

No content.

Description

Audio frequency noise suppression circuit

1 shows an embodiment of a noise suppression circuit according to the present invention.

FIG. 1 is a waveform diagram showing an example of a control signal for suppressing a stereo signal when noise occurs in FIG.

2 shows an embodiment of a residual wave rectifier and a smoothing filter used in FIG.

3 is a circuit diagram showing an embodiment of the converter used in FIG.

4 is a schematic diagram showing an implementation of the noise suppression circuit of the present invention by a digital method.

The present invention relates to a circuit for suppressing audio frequency noise in a signal generated by a record plate.

Document "Electronics Letters", December 9, 1976, Vol 12, No. 25, M.R. Sach and J.M. on pages 656-657. The circuit described in Bullingham's paper "Audio disc scratch filter" includes a signal processing unit disposed between an input terminal and an output terminal having a delay unit and a defective noise suppressor, and a control signal unit connected to the input terminal having a noise detector. .

In the circuit of the prior art, in the absence of noise, the delayed signal in the delay unit is transmitted to the output terminal via a noise suppressor. However, if a pulsed indirect signal, for example a pulsed interference signal due to scratches on the record plate, is detected at the noise detector, a control signal is generated to operate the noise suppressor. Switching is performed at and a non-delayed input signal is applied to the output terminal instead of the delayed input signal. This switching state, that is, the switching state in which the input terminal is connected to the output terminal through the noise suppressor, lasts for 70m sec, and then automatically returns to its original state so that the non-delayed signal is replaced by the delayed signal. However, in such a prior art circuit, when an interference signal is generated during the switching state period, i.e., when the force terminal is connected to the output terminal through the noise suppressor, the interference signal cannot be suppressed, and as a result, There is a drawback to being perceived by hearing.

An object of the present invention is to improve a known circuit.

An audio frequency noise suppression circuit according to the present invention has an input terminal and an output terminal, and includes a delay unit coupled to the input terminal and a noise suppressor coupled in a multiplier form with a control terminal between the delay unit and the output terminal. And a control signal unit including a noise detector coupled to the input terminal and a function generator coupled between the noise detector and a control terminal of the noise suppressor, wherein the function generator generates a control signal and generates a control signal when noise is generated. The signal from the plate is changed from the operating level to the suppression level during the first period, held at the suppression level during the second period, and then changed back from the suppression level to the operating level during the third period. The delay period of the delay unit is equal to the first period.

"Journal of the Audio Engineering Society", October 1068, Vol. 16, No. 4, B.L. on pages 426-429. Cardozo and G.Domburg's paper "An Estimation of Annoyance Caused by Dropouts in magnetically Recorded Music" includes signal interruptions that do not exceed a duration of 10 msec, i.e., the decrease and increase in the amplitude of the signal over any period within the duration. It is known that signal interruption as such is independent of the value of the signal amplitude change and cannot be sensed audibly.

The present invention utilizes such a phenomenon to provide a circuit which artificially generates a signal interruption and replaces the pulsed interference signal existing in the original signal with the signal interruption. In this case, the pulsed interference signal may occur during recording operation due to damage and contamination of the record plate.

As a result of the experiment, the interfering signal is suppressed to the extent that it cannot actually be detected by phonation in the above-described signal reduction period, signal suppression period and signal enhancement period.

In one embodiment of the present invention, the function generator comprises a multivibrator connected to a noise detector, an integrating network for integrating the output pulses of the monostable multivibrator, and a noise for limiting the output pulses of the integrating network to a single amplitude value. And a subordinate connection circuit with an amplitude limiter connected to the control input terminal of the suppressor.

Each time an interference signal is detected by a noise detector, a monostable multivibrator obtains a rectangular pulse with a constant duration and amplitude, which is integrated in the integrating network and then limited by the amplitude limiter to the above-mentioned changes. In response, a control signal of a shape suitable for varying the signal gain is obtained.

In addition, in an embodiment according to the present invention, the integrating network includes an operational amplifier connected in parallel with a capacitor, wherein the previous integrating network is connected to the output terminal of the monostable multivibrator through a resistor, and the output terminal of the operational amplifier. The self-limiting limiter is configured to be coupled to a common connection point of the first and second zener diodes arranged between the power supply and the ground point.

1 shows a signal processor 9 disposed between input terminals 1 and 2 and output terminals 3 and 4, and a control signal unit 10 provided in a forward control loop. The circuit is shown accordingly. The signal processor 9 constitutes two identical parallel signal paths, each of which has a preamplifier (5) subordinated between the input terminals (1) and (2) and the output terminals (3) and (4). And (6), delay circuits 7a and 7b, and controlled amplifiers 8a and 8b having a function as a multiplier. The delay circuits 7a and 7b and the controlled amplifiers 8a and 8b respectively constitute a delay unit 7 and a suppressor 8, respectively.

The control signal section 10 includes a preamplifier 13 and 14, a differential amplifier 29 connected to the output terminals 90 and 91 of the preamplifier 13 and 14, and also a signal. And a summation amplifier 21 'connected to the output terminals 90 and 91 through the resistors 19 and 20 of the addition circuit 21. The differential amplifier 29 is coupled to the differential signal input terminal 12 by the dynamic compressor 11 through the differentiator 53, and the summing amplifier 21 'is the sum signal input terminal of the dynamic compressor 11 ( 15). The output terminal 33 of the dynamic compressor 11 is coupled to the monostable multivibrator 36 via amplifiers 30 and 35. The monostable multi-vibrator 36 is connected to the converter 37, which converts the pulse frequency into a DC voltage. The threshold voltage is set in the pulsator 30.

The converter 37 is connected in parallel with the coupling capacitor 38 and the resistor 39 connected in series on the cutting line 17 (along with the amplifiers 30, 35 and the monostable multivibrator 36). The output 33 of the dynamic compressor 11 is coupled to the input 40 of the noise detector 32 via cut line 17, and also the transformer 37 is coupled to the input terminal 40. The output 18 of the noise detector 32 is coupled to the function generator 34, and the output of the function generator 34 is connected to the control input terminal 49 of the noise suppressor 8.

The left and right audio signals L and R of the audio frequency stereo signal recorded on the record plate are separately applied to the input terminals 1 and 2. In the signal processing section 9, the audio signals are amplified in the same range by the preamplifiers 5 and 6, are delayed over an equal period in the delay unit 7, and noise is generated as a result of scratches on the record. In the case of a row, the noise is suppressed in the noise suppressor 8 by suppressing the signal. The control signal for suppressing the signal is applied to the noise suppressor 8 through the control signal input terminal 49, which controls the noise in the audio signals L and R applied to the input terminals (1) and (2). When detected, the control signal unit 10 is applied to the control terminal 49 of the noise suppressor 8.

In addition to the amplitude change due to the dynamic range, the jagged stereo signal has a fill-like amplitude change that is superimposed on it and shorter in duration than the change in amplitude. These amplitude changes may be caused by scratch noise on the record plate, which in this case is undesirable. However, these amplitude changes may also be generated from instruments such as trumpets or percussion instruments, in which case the amplitude changes are desirable. The desired pulsed amplitude change is a music pulse, and the pulsed amplitude change of the fluorine network will be recovered later in the specification as a noise pulse.

Noise pulses are detected at both stages.

In the first selection stage, the pulsed amplitude change, that is, the noise pulse and the music pulse, are separated from the amplitude change caused by the dynamic range. Also, in the second select stage, the noise pulse is separated from the music pulse. To achieve this goal, we use the fact that the time interval between noise pulses is larger than the time interval between music pulses.

The first selection stage may be realized by an amplifier having a threshold that only amplifies a signal having an amplitude exceeding a predetermined threshold voltage. This threshold voltage can be changed to some extent with the dynamic range of the stereo signal to exceed only the superimposed noise and music pulses above the threshold voltage value. Alternatively, a threshold voltage may be set to a fixed value to gently compress the disturbed stereo signal to some extent, such that only superimposed noise and music pulses in the dynamic range are exceeded.

The circuit as shown in FIG. 1 forms the difference signal L-R in the differential amplifier using the scheme described below for the first selection stage. The reason is the publication "Electronic Letters" published December 6, 1976, Vol. 12, M.R. 656-657. Sach and J.M. As can be seen from Bullingham's paper "Audio disc Soratch fillter," the difference signal L-R of the stereo signal from the record plate is particularly suitable for the detection of noise caused by scratches on the record plate concerned.

In the circuit of this embodiment, the difference signal L-R is preferentially differentiated in the differentiator 53. As a result, the amplitude of the music and noise pulses is increased with respect to the amplitude change caused by the dynamic range of the stereo signal. For this purpose, the differentiator is actually given a 0.14 msec. The difference signal L-R is then compressed in the controllable amplifier 16a of the dynamic compressor 11. At this time, the control signal necessary for compression should represent the dynamic range of the undisturbed stereo signal. The approximate signal that can be tolerated for the dynamic range of the undisturbed stereo signal is the sum signal LR, for which the noise pulse in the sum signal is extremely small compared to the rest of the music signals in the respective signals L, R and LR. Can be mentioned.

The sum signal L R is formed in the signal addition circuit 21 and applied to the controllable amplifier 16b of the dynamic compressor 11, which is a high negative orbital amplifier. The negative feedback of the controllable amplifier 16 is performed through the amplifier stage 23, the full-wave rectifier 24, and the smoothing pin 25 which are cascaded between the output terminal 22 and the control input terminal 28. The smoothing filter operates as an RC control network. By timely selecting the time constant of the smoothing filter 25 to be large, it is possible to perform the pulse type amplitude change, that is, the noise and the music pulse not to be followed. As a result, the control signal applied to the control input terminal 28 accurately represents the dynamic range of the stereo signal, and when the amplitude of the sum signal is increased or decreased, the two amplifiers 16a and 16b are controlled to reduce the gain. Or increase.

As a result, both the strong and weak components of the difference signal L-R are actually adjusted to the same amplitude level.

Since noise and music pulses are not actually present in the control signal, they are not removed from the difference signal L-R and remain.

After this dynamic range of compression, the difference signal L-R is passed through an amplifier 30 having a threshold, where the first selection of music and noise pulses is made from the difference signal L-R. For this purpose, the threshold voltage of the amplifier 30 is selected such that only music and noise pulses are exceeded above the threshold voltage. These music and noise pulses are amplified by the amplifier 35 and then operate the monostable multivbrator 36.

The second choice, i.e., discrimination between noise and music pulses, is generally based on the fact that noise pulses occur independently, and that music pulses occur in five or more series of pulses having a frequency of about 120 Hz or more. Therefore, the pulse frequency of the output pulse of the monostable multivibrator 36 includes information on the pulse shape. This pulse frequency is converted into a negative DC voltage in the converter 37. In practice, by appropriately selecting the charging time constant of the converter 37, the DC voltage is applied to a pulse train including 5 or more pulses within a 20 msec period. By properly selecting the discharge time constant of the converter 37, the DC voltage decreases exponentially from the maximum to the minimum when no pulse is generated for 200 msec.

Like the first selection, the second selection can be realized by an amplifier having a threshold for amplifying only a signal having an amplitude exceeding a predetermined threshold voltage. The threshold voltage may be set to a predetermined fixed value. In this case, by using the DC voltage of the converter 37 as a control voltage, first, a noise pulse present in the output signal of the dynamic compressor 11 or the monostable multivibrator 36 is present in the output signal in the controllable amplifier. After increasing for the music pulses, these noise and output pulses are sent to an amplifier with a fixed threshold. In this case, the threshold voltage of the amplifier having a threshold value needs to be selected to have an amplitude value of the noise pulse and an amplitude value of the music pulse.

Alternatively, the threshold voltage may be changed together with the DC voltage of the transformer 37 to directly apply noise and music pulses generated in the output signal of the dynamic compressor 11 or the monostable multivibrator to an amplifier having a variable threshold. This selection method is used in the circuit of this embodiment. The direct current voltage of the converter 36 is applied to the base of the transistor 41 which is included in the noise detector 32 via the input terminal of the noise detector 37 and acts as an amplifier with a threshold. In addition, the difference signal L-R which is an output signal of the dynamic compressor 11 is also applied to the input terminal 40 through the wiring 17. In this case, the emitter of the transistor 41 is grounded and the collector is connected to the collector resistor and the output terminal 18.

The threshold voltage of transistor 41 is determined by it's base-emitter bias voltage. This voltage increases when the value of the DC voltage portion of the converter increases in the negative direction. The music pulses applied to the base of the transistor 41 through the wiring 17 generate a negative DC voltage at the output terminal of the converter which cannot conduct the transistor 41. On the contrary, the noise pulses do not generate negative DC voltage at the output terminal of the converter so that the transistor 41 can be conducted by the noise pulse which has lost the wiring.

However, noise pulses may occur in the heat of the music pulses, which are recognized as the desired music pulses because they do not conduct the transistor 41. In practice, these noise pulses can hardly be perceived in the associated music signal. The noise pulse generated at the output terminal 18 of the noise detector 32 starts the monostable multivibrator 43 of the function generator 34.

The square wave output pulse of the monostable multivibrator 43 is an integrating network composed of a parallel connection of a variable resistor 44 connected to the multivibrator 43 and an operational amplifier 45 and a capacitor 46 connected in series thereto. 50) and integrated into a triangular pulse. The amplitudes of these triangular pulses are limited in the limiter circuit 51 of the function generator 34, which consists of the zener diodes 47 and 48 in a series connection arrangement, which are common to these zener diodes. The connection point is connected to the integrated circuit 50 and the control terminal 49 of the noise suppressor 8. FIG. 1A shows the shape of the control pulse, and the time passing through the paths a, b and c is about 1 to 2 msec. During path b, the stereo signal is suppressed to the maximum at the noise suppressor 8. During paths a and b, the gain of the stereo signal is reduced and amplified respectively in the controlled amplifiers 8a and 8b of the noise suppressor 8, respectively.

The delay period in the delay unit 7 needs to be at least equal to the period from the time when the noise is detected until the path b of the induced control pulse due to this occurs.

In the circuit shown in FIG. 1, amplifiers 5, 6, 13, 14, 21 '29, 35 and 45 are operational amplifiers of the type TCA 680. The amplifier 30 is a transistor of the BC 550 type, the monostable multivibrators 36 and 43 are provided as an integrated circuit of the HEF 4528 type. The delay unit 7 is configured in the form of a TDA 1022 and realizes a signal delay of 3 to 1.5 msec per channel under the control of the clock frequency which can be adjusted between 85 KHz and 170 KHz. Noise suppressor 8, control amplifiers 16a and 16b are of type TCA 730 and transistor 41 is of type BC 550. A pulse width of 2 msec is selected for the monostable multivibrator 36 and a pulse pole of 5 msec is selected for the monostable multivibrator 43.

FIG. 2 is a detailed view of the preferred embodiment of the full-wave rectifier 24 and the smoothing filter 25, and the same reference numerals are given to the corresponding elements in FIG.

FIG. 2 is a detailed view of the preferred embodiment of the full-wave rectifier 24 and the smoothing filter 25, and the same reference numerals are given to the corresponding elements in FIG.

Full-wave rectifier 24 includes operational amplifiers 61 and 62, and the non-inverting and inverting input terminals of the operational amplifiers 61 and 62 input signals through matching resistors 63 and 64. It is connected to the terminal 31, respectively. At this time, the negative passages of the operational amplifiers 61 and 62 are respectively formed as paths from their outputs to the inverting input terminals through the diodes 65 and 67, respectively. The cathodes of diodes 65 and 67 are connected to the inverting input terminals and are also connected to the input terminal 71 of smoothing filter 25 via resistors 69 and 70. The output terminals of operational amplifiers 61 and 62 are also connected to the cathodes of diodes 66 and 68, respectively, and are coupled to input terminal 71 via diodes 66 and 68, respectively. . The non-inverting input terminal of the operational amplifier 62 is also connected to the reference voltage point.

The smoothing filter 25 includes a resistor 26 and a smoothing capacitor 27 connected in series between the input terminal 71 and ground. At this time, the connection point between the resistor 26 and the smoothing capacitor 27 is connected to the control input terminal 28.

When the signal polarity of the input terminal 31 is positive with respect to the reference voltage of the non-inverting input terminal of the operational amplifier 62, the diode 65 is conducted so that the negative feedback of the operational amplifier 61 is greatly enlarged. The gain is negligibly small. In addition, this positive signal is amplified and inverted and appears at the output of the operational amplifier 62. In this case, since the diode 67 is in a blocking state and the diode 68 is in a conductive state, Appears on the input terminal 71. In this case, since the diode 66 is blocked, the output impedance of the operational amplifier 61 is not loaded on the output of the operational amplifier 62.

When the signal polarity of the input terminal 31 is negative with respect to the reference voltage, the diode 65 is cut off and the diode 66 is conducted. At this time, the output signal of the operational amplifier 61 is amplified and appears on the input terminal 71 of the smoothing filter 25 through the diode 66 in the conductive state. In this case, since the diode 67 is in the conduction state, the negative feedback of the operational amplifier 62 is greatly enlarged so that the gain is negligible. In this case, since the diode 68 is in the cutoff state, the output impedance of the operational amplifier 62 is not loaded on the output of the operational amplifier 61.

Therefore, a signal having the same amplitude as the difference component between the reference voltage of the non-inverting input terminal of the operational amplifier 62 and the amplitude of the sum signal L + R of full-wave rectification appears on the input terminal 71 of the smoothing filter 25. Is increased when the amplitude of the sum signal decreases, and decreases when it increases. This signal is smoothed by the smoothing filter 25 and flows out as a control signal for the controllable amplifiers 16a and 16b. In this case, the smoothing filter 25 is given a time constant of about 3.3 msec. In this time constant case, the control signal represents only the amplitude change caused by the dynamic range of the stereo signal and does not actually represent noise and music pulses.

In practice, operational amplifiers 61 and 62 are comprised of TCA 680 integrated circuits, the diodes being in the form of BAX 13, with a reference voltage of 8 volts. The values of the resistors 63, 64, 69, and 70 are 10 K ?, and the capacity of the smoothing capacitor 27 is 1 k ?.

3 is a detailed view of a preferred embodiment of a transducer disposed between the output terminal 80 of the monostable multivibrator 36 and the input terminal 40 of the noise detector 32.

The converter 37 comprises a coupling capacitor 81, a resistor 83 and a diode 84 connected in series, the cathode of which is connected to a resistor 83. The anwood of diode 84 is also grounded through an RC parallel circuit consisting of capacitor 85 and potentiometer 86. The mover of the potentiometer 86 is coupled to the input terminal 41 of the noise detector 32 through a matching resistor 87, and the connection point between the coupling capacitor 81 and the resistor 83 is connected through the matching resistor 82. Grounded.

Negative pulses present on the output terminal 80 of the monostable multivibrator 36 charge the capacitor 85 negatively through a diode 84 that is conductive to these pulses. This capacitor 85 is a potentiometer Discharged through the body of 86, in practice, the time constant is about 0.1 sec. At this time, if the rated value of the RC element is appropriately selected, a negative maximum voltage of about -2.2 V is obtained in the capacitor 85 with respect to a pulse train including 5 to 6 pulses within 20 mseb. The voltage adjustable portion of the capacitor 85 is derived by the mover of the potentiometer 86 and applied to the input terminal 40 of the noise detector 32 via the matching resistor 84 as a variable threshold voltage. In this case, if the operating difference is properly adjusted, the noise detector has a maximum sensitivity at this minimum value if the threshold voltage of the noise detector reaches a minimum value after no pulse is generated for 20 msec.

In practice, capacitors 81 and 85 have capacitance values of 22 kHz and 1 kHz, respectively, and resistors 82, 83 and 87 have values of 39 kHz, 5.1 kHz and 39 kHz, respectively. Each one has it. Potentiometer 86 has a maximum value of 100 Hz. The diode 84 is of the BAX 13 type.

FIG. 4 shows another embodiment of the noise suppression circuit according to the present invention. In this example, noise suppression is performed digitally on a pulse code modulation (PCM) signal representing an audio frequency stereo signal. Denotes the same reference number.

The digital signal is applied to the input terminal 1. By way of example, this digital signal may be configured as a sequence of digital codewords representing a numerical value corresponding to the sampling value of the analog audio frequency signal. In the case of a stereo signal, these codewords may represent sampling values of left and right audio signals.

In such a digital signal, in order to suppress that an interference signal that can be detected after digital / analog conversion is not detected, in the present invention, when a corresponding interference signal is detected, a codeword in a multiplier 8 operating as a noise suppressor is operated. Multiplying the weighting coefficient by a decrease from " 1 " to the minimum value " 0 " for a period of 1 and keeping the minimum value " 0 " for a second period and increasing to " 1 "

This weighting coefficient can simply be generated in the function generator 34 via a ROM which is read under the control of the clock signal for each interference signal.

For simple detection of disturbed codewords, one or a plurality of check bits are added to each codeword during protection of the sample value. The value of these check bits is determined by the value of the data bits, for example by a simple way of keeping the total number of " 1 " in each codeword and associated check bits constant.

Thus, by counting the number of " 1 " in each codeword and the associated check bit in the adder 32, which acts as a noise detector, it is possible to recognize the disturbed codeword. If the number of consecutive perturbed codewords exceeds a predetermined number and begins to be detected as noise after the digital-analog conversion, the function generator 34 is activated and the noise suppression according to the invention is carried out in the multiplier 8. do.

The delay unit 7 may be equipped with a shift register in a known manner. In that case, the delay time needs to be at least equal to the period during which codewords exceeding the predetermined number are generated and the first period.

Claims (1)

Audio frequency noise in the signal generated from the record plate, comprising a signal processing unit disposed between the input terminal and the output terminal by installing a delay unit and a noise suppressor coupled thereto, and a control signal unit coupled to the input terminal by installing a noise detector. In the audio frequency noise suppression circuit for suppressing the noise suppression, the noise suppressor 8 is disposed between the delay unit 7 and the output terminals 3 and 4 in the form of a multiplier, and the control input of the multiplier-type noise suppressor 8 is controlled. An audio frequency noise suppression circuit characterized in that a terminal (49) is connected to a noise detector (32) via a function generator (34) for generating a control signal.

Images